Introduction:
Oracle Restart is a component added to the 11gR2 in order to improve the high availability of different database resources.
Database instances, listeners are resources that can be automatically restarted after a hardware or software failure or whenever the server host restarts.
These components can be easily added to the Oracle Restart using the srvctl add command.
The procedure of adding the Enterprise Manager database control as a resource is different. In this article, I am going to show you how to add this component to the Oracle Restart.
By adding Enterprise Manager database control to the Oracle Restart, the resource will be automatically restarted with the whole cluster resource pack.
Test Environment: The following procedure has been tested on RAC 6 nodes 11.2.0.2.2 version under Redhat Enterprise Linux.
Remember, the database version must be from 11gR2.
The procedure:
Step1: With your gird user, create 2 scripts under $GRID_HOME/crs/script (GRID_HOME or CRS_HOME), in my case: /opt/11.2.0/grid/crs/script/.
See the content of those 2 scripts below;
-bash-3.2$ cat crs_dbconsole.sh
#!/bin/bash
#
# dbcagent.sh - script to start and stop the dbconsole 11gR2 agent
#
# description: Oracle 11gR2 database console agent
ORACLE_BASE=/opt/app/oracle
ORACLE_HOME=/opt/app/oracle/product/11.2.0/db_1
LIBRARY_PATH=$ORACLE_HOME/lib
ORACLE_SID=db01
ORACLE_UNQNAME=racdb
PATH=$ORACLE_HOME/bin:$PATH
export ORACLE_BASE
export ORACLE_HOME
export LD_LIBRARY_PATH
export ORACLE_SID
export ORACLE_UNQNAME
export PATH
agent_start () {
sudo -u oracle /opt/11.2.0/grid/crs/script/agentstartup.sh start
}
agent_stop () {
sudo -u oracle /opt/11.2.0/grid/crs/script/agentstartup.sh stop
}
agent_check () {
sudo -u oracle /opt/11.2.0/grid/crs/script/agentstartup.sh check
}
case "$1" in
start)
agent_start
;;
stop)
agent_stop
;;
check)
agent_check
;;
clean)
agent_clean $(ps -C emagent -o pid=)
;;
*)
echo $"Usage: `basename $0` {start|stop|status|clean}"
exit 1
esac
-bash-3.2$
-bash-3.2$ cat agentstartup.sh
#!/bin/bash
#
# dbcagent.sh - script to start and stop the dbconsole 11gR2 agent
#
# description: Oracle 11gR2 database console agent
ORACLE_BASE=/opt/app/oracle
ORACLE_HOME=/opt/app/oracle/product/11.2.0/db_1
LIBRARY_PATH=$ORACLE_HOME/lib
ORACLE_SID=db01
ORACLE_UNQNAME=racdb
export ORACLE_BASE
export ORACLE_HOME
export LD_LIBRARY_PATH
export ORACLE_SID
export ORACLE_UNQNAME
agent_start () {
$ORACLE_HOME/bin/emctl start dbconsole
}
agent_stop () {
$ORACLE_HOME/bin/emctl stop dbconsole
}
agent_check () {
$ORACLE_HOME/bin/emctl status dbconsole
}
case "$1" in
start)
agent_start
;;
stop)
agent_stop
;;
check)
agent_check
;;
clean)
agent_clean $(ps -C emagent -o pid=)
;;
*)
echo $"Usage: `basename $0` {start|stop|status|clean}"
exit 1
esac
-bash-3.2$
Step2: Give the appropriate privileged to the 2 scripts created in the step1;
[root@racdbdb01 script]# cd /opt/11.2.0/grid/crs/script [root@racdbdb01 script]# chown grid:oinstall /opt/11.2.0/grid/crs/script/* [root@racdbdb01 script]# chmod 755 * [root@racdbdb01 script]# ls -rtl total 8 -rwxr-xr-x 1 grid oinstall 951 Sep 16 12:19 crs_dbconsole.sh -rwxr-xr-x 1 grid oinstall 838 Sep 16 12:20 agentstartup.sh [root@racdbdb01 script]#
Step3:
As a root user, edit the /etc/sudoers file; do it for all nodes of your cluster:
First, change the file permission from 440 to 640. Then, edit the sudoer file like described below;
Finally, do not forget to back the permissions to 440 on this sudoer file.
[root@racdbdb01 script]# chmod 640 /etc/sudoers [root@racdbdb01 script]# vi /etc/sudoers add at the end of the file; grid racdbdb01 =(oracle) NOPASSWD: /opt/11.2.0/grid/crs/script/agentstartup.sh start grid racdbdb01 =(oracle) NOPASSWD: /opt/11.2.0/grid/crs/script/agentstartup.sh stop grid racdbdb01 =(oracle) NOPASSWD: /opt/11.2.0/grid/crs/script/agentstartup.sh check and comment out: #Defaults requiretty [root@racdbdb01 script]# chmod 440 /etc/sudoers [root@racdbdb01 script]#
step4: As oracle user, change the permissions of the following files. Do it for all nodes in your cluster.
-bash-3.2$ su - oracle Password: -bash-3.2$ chmod 664 /opt/app/oracle/product/11.2.0/db_1/sysman/config/emd.properties -bash-3.2$ chmod 664 /opt/app/oracle/product/11.2.0/db_1/racdbdb01_ddb/sysman/config/emd.properties -bash-3.2$ chmod 755 /opt/app/oracle/product/11.2.0/db_1/bin/emctl -bash-3.2$ -bash-3.2$
step5: Copy agentstartup.sh and crs_dbconsole.sh files to all remaining nodes of the cluster.
Do not forget to replace the SID values by the SID of each instance of your RAC:
-bash-3.2$ scp agentstartup.sh root@racdbdb06:/opt/11.2.0/grid/crs/script/agentstartup.sh root@racdbdb06's password: agentstartup.sh 100% 838 0.8KB/s 00:00 -bash-3.2$ -bash-3.2$ scp crs_dbconsole.sh root@racdbdb03:/opt/11.2.0/grid/crs/script/crs_dbconsole.sh root@racdbdb03's password: crs_dbconsole.sh 100% 951 0.9KB/s 00:00 -bash-3.2$
Step6: Add a new resource “crs_dbconsole” to the Oracle Restart.
From only node 1 and as grid user:
-bash-3.2$ crsctl add resource crs_dbconsole -type local_resource -attr " ACTION_SCRIPT=/opt/11.2.0/grid/crs/script/crs_dbconsole.sh, DESCRIPTION=Local DB Console Ressource for DB Console Agent, DEGREE=1, ENABLED=1, AUTO_START=restore, START_TIMEOUT=120, UPTIME_THRESHOLD=1h, CHECK_INTERVAL=60, STOP_TIMEOUT=120, SCRIPT_TIMEOUT=120, RESTART_ATTEMPTS=3, OFFLINE_CHECK_INTERVAL=60, START_DEPENDENCIES=, STOP_DEPENDENCIES=,"
Step7: test by starting and stopping the new added resource;
-bash-3.2$ crsctl start res crs_dbconsole CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb03' CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb05' CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb01' CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb06' CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb02' CRS-2672: Attempting to start 'crs_dbconsole' on 'racdbdb04' CRS-2676: Start of 'crs_dbconsole' on 'racdbdb05' succeeded CRS-2676: Start of 'crs_dbconsole' on 'racdbdb02' succeeded CRS-2676: Start of 'crs_dbconsole' on 'racdbdb01' succeeded CRS-2676: Start of 'crs_dbconsole' on 'racdbdb06' succeeded CRS-2676: Start of 'crs_dbconsole' on 'racdbdb04' succeeded CRS-2676: Start of 'crs_dbconsole' on 'racdbdb03' succeeded -bash-3.2$ -bash-3.2$ crsctl stop res crs_dbconsole CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb04' CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb05' CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb01' CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb03' CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb06' CRS-2673: Attempting to stop 'crs_dbconsole' on 'racdbdb02' CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb05' succeeded CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb04' succeeded CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb03' succeeded CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb02' succeeded CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb06' succeeded CRS-2677: Stop of 'crs_dbconsole' on 'racdbdb01' succeeded
Let ‘s reboot one of the servers for more testing, Reboot for example the node 2 of the cluster:
-bash-3.2$ ssh root@racdbdb02 root@racdbdb02's password: Last login: Fri Sep 16 12:56:07 2011 from racdbdb01.ddomain.com [root@racdbdb02 ~]# [root@racdbdb02 ~]# init 6 [root@racdbdb02 ~]#
Wait for some minutes and then check the resource:
-bash-3.2$ crsctl status res crs_dbconsole NAME=crs_dbconsole TYPE=local_resource TARGET=ONLINE , ONLINE , ONLINE , ONLINE , ONLINE , ONLINE STATE=ONLINE on racdbdb01, ONLINE on racdbdb02, ONLINE on racdbdb03, ONLINE on racdbdb04, ONLINE on racdbdb05, ONLINE on racdbdb06 -bash-3.2$ -bash-3.2$ su - oracle Password: -bash-3.2$ emctl status dbconsole Oracle Enterprise Manager 11g Database Control Release 11.2.0.2.0 Copyright (c) 1996, 2010 Oracle Corporation. All rights reserved. https://racdbdb01.ddomain.com:1158/em/console/aboutApplication EM Daemon is running. ------------------------------------------------------------------ Logs are generated in directory /opt/app/oracle/product/11.2.0/db_1/racdbdb02_ddb/sysman/log -bash-3.2$
Conclusion:
Enterprise Manager database control daemon will automatically restart after a hardware or software failure or whenever the server host restarts.
Scridb filter
Blog's RSS Feed
Loading ...
[...] 11gR2 RAC: add Enterprise Manager Database Control to crs resources for automatic restart [...]